Research On Condition Monitoring And Fault Diagnosis System For Centrifugal Pump Unit Driven By Digital Twin

In the traditional condition monitoring and fault diagnosis process,the degree of digitization and intelligence of equipment is not high enough to realize real-time mapping between physical space and information space,real-time fault prediction,and fault models synchronization,resulting in slow fault feedback,low operation and maintenance efficiency,and high maintenance costs.In this regard,this thesis proposes a new scheme for the condition monitoring and fault diagnosis of the centrifugal pump unit driven by the digital twin.Based on the new scheme,key technologies such as the construction of digital twin mapping model,real-time fault diagnosis,diagnosis result verification,digital twin model correction,and deep learning model reconstruction have been deeply studied.The main contents of this thesis are as follows:(1)The overall scheme design of the condition monitoring and fault diagnosis system for centrifugal pump unit.Based on the demand analysis of the condition monitoring and fault diagnosis process,the corresponding research goals and overall plan are formulated for the system,and the architecture and function modules of condition monitoring and fault diagnosis system for centrifugal pump unit driven by digital twin are designed.(2)Research on construction technology of digital twin mapping model for centrifugal pump unit.By constructing the digital twin mapping model,the digital twin technology is integrated into the health management process to realize the service functions of condition monitoring,fault diagnosis and fault maintenance for the centrifugal pump unit.Based on the entity of the centrifugal pump unit,the data collection method,twin model construction method,twin data construction method,connection method,and system service function module integration are discussed accordingly.(3)Fault diagnosis technology based on data driven and model driven.The study of intelligent diagnosis technology is deeply conducted based on the digital twin mapping model.Based on the collected real-time vibration data,the fault diagnosis method of deep learning is used to analyze and predict the fault for the centrifugal pump unit in real time.According to the intelligent diagnosis results,the model driven fault diagnosis method is used to analyze the reliability of the results.After the fault is determined,the twin model of centrifugal pump unit is modified to ensure the consistency of physical entity and virtual model,so as to realize the synchronization of fault model.In order to ensure the accuracy of fault diagnosis,the model reconstruction of deep learning is realized by constructing the corresponding feedback mechanism based on the driving condition of verification results.(4)Realization of condition monitoring and fault diagnosis system for centrifugal pump unit driven by digital twin.Based on the design of system architecture and functional modules and the research of key technologies,the operation mechanism,process and human-computer interface of the system are designed,and the prototype system is developed.Finally,the feasibility and stability of the system are verified through three different working conditions.